Abstract
While the number of legalized recreational marijuana markets continues to grow in the USA, state and local governments are still determining how best to levy taxes on marijuana receipts in the face of consumer behavioral responses, such as stockpiling behavior and cross-border purchasing. Using the introduction of a 25% tax on marijuana in Oregon as a natural experiment, we conduct difference-in-differences, regression discontinuity, and event study analyses to identify the effect of the tax increase on marijuana prices and quantities, consumer stockpiling, and cross-border purchasing. Our results are consistent with the theoretical predictions of tax incidence—finding that consumer marijuana prices rise and the quantity of marijuana sold falls as a result of the tax. We also observe evidence of short-term stockpiling of marijuana in anticipation of the tax and find that the tax led to increased cross-border substitution, most notably, at the Washington State border. Based on our results, we also determine that supply is relatively elastic in this market, and demand becomes more elastic over time as consumers gain greater information on prices of substitutes.
Similar content being viewed by others
Notes
A Schedule I controlled substance is a drug, substance, or chemical that: has a high potential for abuse; has no currently accepted medical value; and has no ability to be used or tested safely, even under medical supervision; and is subject to regulatory controls and administrative, civil, and criminal penalties under the Controlled Substances Act of 1970.
According to Gallup (http://www.gallup.com/poll/196550/support-legal-marijuana.aspx), 60% of adults in the United State supported legalization of marijuana, up from 35% in 2003/2005. The highest public support for marijuana legalization is from 18 to 34-year olds, of which 77% support legalization. The Pew Research Center (http://www.pewresearch.org/fact-tank/2016/10/12/support-for-marijuana-legalization-continues-to-rise/) finds similar levels of support, with 57% of adults and 71% of Millennials in support of marijuana legalization in 2016.
Washington, Oregon, Colorado, Alaska, California, Massachusetts, Maine, Nevada, Vermont, and Michigan.
According to the Oregon Department of Revenue, the marijuana tax revenue would be disbursed as follows: 40% would go toward funding education, 20% would go to Mental Health, Alcoholism, and Drug Services, 15% would go to state law enforcement, 10% each to cities and counties for marijuana enforcement costs, and 5% to the Oregon Health Authority for alcohol and drug abuse prevention (Oregon Department of Revenue).
Additional local option taxes of 3% or less may also be levied on recreational marijuana sales if approved by voters (House Bill 3400). These local option taxes were first voted on in the November 2016 election and thus fall outside of the sample period of this study.
As of this writing, California, Colorado, Maine, Massachusetts, Nevada, Oregon, and Washington have a tax on marijuana. Washington currently has the highest tax rate at 37%, while Oregon is the second highest at 17%. Massachusetts has the lowest tax rate at 3.75%. Colorado and Maine have a 10% tax rate, while Nevada has a 15% tax rate. Massachusetts and Oregon also allow local governments to levy an additional 2 or 3% tax on marijuana, respectively. For a full history of these legislations, see Appendix Table 5.
Laws have been established that dictate how much individuals are legally able to stockpile. In Oregon, individuals may possess eight ounces (~ 226 g) of marijuana in their homes at any given time and may grow up to four plants. Individuals who are found to possess one pound or more of marijuana are at risk of serving anywhere from 6 months to 1 year in prison.
Since marijuana is federally classified as a Schedule I controlled substance, transporting marijuana across state lines or stockpiling large amounts of marijuana could lead to drug trafficking charges, with penalties ranging from up to 5 years in federal prison and a fine of $250,000 to $1 million for the 1st offense and up to 10 years in federal prison and a fine of $500,000 to $2 million for the 2nd offense. Oregon has also passed legislation (House Bill 4014) to criminalize the importing or exporting of marijuana from/into other states. While less than an ounce of marijuana yields a $260 fine, holding more than an ounce is a misdemeanor carrying a $6250 fine.
For a discussion of the incidence of a sales tax without stockpiling in competitive, monopoly, and oligopoly markets, see Myles (1995), Fullerton and Metcalf (2002), and Tremblay and Tremblay (2017). For a discussion of the effects of stockpiling and cross-border substitution on consumption and the tax incidence in competitive markets, see an earlier version of this working paper, which is available upon request.
It is key to note that Hansen et al. (2017b) assess the degree of cross-border substitution response to the introduction of a tax in an adjacent state. They do not, however, quantify these effects for the state implementing the tax. Thus, our results provide a comparison to and complement the results of their analysis.
Fairbairn et al. (1976) conclude that herbal or resin cannabis or extracts are reasonably stable for 1 or 2 years if stored in a dark room at room temperature. The same study finds that it can remain effective and safe to consume for much longer as long as the essential oils slowly break down over time. A more recent study by Confidence Analytics in Washington State investigated the shelf life of marijuana (www.marijuanaventure.com/cannabis-shelf-life/). They examined the decay of THC levels over time and found that the loss in potency over 125 days was very minimal. In fact, the results suggest that when stored completely dried, marijuana would retain half of its potency after a decade of storage. However, if ideal storage conditions are not followed, mold and other issues may reduce shelf life considerably.
For a theoretical discussion of stockpiling and cross-border substitution in these types of markets, see the earlier working paper version of this study. This working paper version is available upon request.
“Smokable marijuana” consists of only the dried buds and leaves that one would smoke. All other forms require processing of some sort, which extracts the active ingredients from marijuana and turns it into other products, commonly referred as “consumable marijuana.” The consumable category includes a wide range of products. For example, cannabis can be infused into bottled drinks, candies, and baked goods. Topical creams can also include cannabis, which does not get one "high" but is used as a pain reliever.
The Washington data set does not include a measure of the number of dispensaries from which the county-level data are aggregated.
We take this tax-exclusive firm revenue and divide by one minus the tax rate in each state to obtain the tax-inclusive total revenue collected per day. Given that there is no tax in Oregon in the pre-tax period, the tax-exclusive revenue and tax-inclusive revenue is the same. In the post-tax period in Oregon, the tax-inclusive revenue is equal to the firm revenue divided by 0.75 (since the tax rate jumped to 25%). The tax rate in both time periods is 37% in Washington and thus, the tax-inclusive revenue in Washington is equal to the firm revenue divided by 0.63.
As noted in Hansen et al. (2017a), there are multiple issues with the raw price data in Washington. See the Hansen, Miller, and Weber (2017a) online data Appendix (https://static1.squarespace.com/static/56a1484625981dd79f45da68/t/59713b6246c3c4fdedfb1b00/1500592994649/data-cleaning-appendix.pdf) for a discussion of how prices are adjusted in that study to deal with these inconsistencies. Due to the aggregate nature of our data, average total revenue, which is our measure of consumer prices, there is little adjustment that can be made to the raw data to deal with some of the inconsistencies identified by Hansen, Miller, and Weber (2017a). Thus, the prices for Washington used here—where we factor in the 37% tax on top of the aggregate firm revenue data we collect—may be considered a lower bound on the actual prices facing consumers in this market.
Although possession and consumption of usable recreational marijuana is legal in Washington, home cultivation of marijuana is illegal in Washington. To grow marijuana, an individual must be a medical marijuana patient (may grow up to six plants) or be a I-502 license holder. I-502 license holders are growers who grow marijuana for purchase from recreational retailers.
Similar trends are found in the sale of marijuana plants and seeds in Oregon and consumable marijuana in Washington. Descriptive statistics for these types of sales are available upon request.
Senate Bill 1601 declared that retail sales to medical marijuana cardholders or designated primary caregivers purchasing on behalf of cardholders were not subject to taxation from the state, cities, or counties.
The effective tax rate in Washington was approximately 37% throughout the period from October 1, 2015 to March 31, 2016. There were also no other concurrent changes in other “sin taxes” in these states at the time the marijuana tax was introduced in Oregon. Cigarette taxes in Oregon increased by one cent (0.76% increase) on January 1, 2016 while cigarette taxes in Washington have been unchanged since May 1, 2010 (https://www.tobaccofreekids.org/assets/factsheets/0275.pdf). There were no changes to tax rates on beer, wine, and spirits in Oregon or at the federal level between 2015 and 2017 (Oregon Liquor Control Commission). While taxes on beer and wine were unchanged in Washington at the start of 2016, taxes on spirits in Washington fell by $1.68 (4.77% decrease) on January 1, 2016 (Tax Foundation). Despite this decrease in taxes on spirits, Washington remained the state with the highest tax rate on spirits.
Given the close proximity of Oregon and Washington, there is some concern that cross-border substitution may bias the difference-in-differences results. Thus, we restrict our control group to counties in Washington that do not share a border with Oregon. This also allows us to use counties in Washington that share a border with Oregon as an additional treatment group in some of our cross-border analyses.
Given that we are using state-level aggregate data, we use heteroscedasticity-robust standard errors. Given we only have two states, we are unable to cluster standard errors at the state-level. Also, data limitations in Oregon do not allow us to cluster the standard errors at lower levels of aggregation (e.g., county or city-level).
Consumer incidence = consumer price rise/size of tax = 3.16/(3.16 + 0.26) = 3.16/3.42 = 0.924
Consumer price rise in percentage terms = size of tax * consumer incidence = 25 * 0.924 = 23.1%.
Hansen et al. (2017b) find that supply in the Washington market is also very elastic at the time of the introduction of the Oregon market, providing further evidence for the similarities between the two markets during this time period.
As noted in Hansen et al. (2017b), these difference-in-difference results might be somewhat inflated, as they find no statistically significant change in sales in Washington border counties after the introduction of the Oregon tax.
This analysis is estimated using the rdrobust command (Calonico et al. 2014b) in STATA 15.
The baseline specification uses a quadratic form of this polynomial.
We also conduct a set of sensitivity analyses using higher order polynomials (3rd, 4th, or 5th order) and a set of sensitivity analyses using alternative bandwidths (30, 60, or 90-day windows). Regardless of the specifications used, the results are largely consistent—statistically significant effects on consumer prices (sizes between $3.18 and $3.29) and insignificant impacts on producer prices and marijuana sales. These results are presented in Appendix Table 7.
These elasticities found in the difference-in-differences (− 0.467) and regression discontinuity (− 0.528) analyses fall within the range found in previous literature on the elasticity of demand of marijuana, which tends to range between − 0.1 and − 0.81 in marijuana markets (see Clements and Zhao 2009; Clements et al. 2010; Davis et al. 2016; Hansen et al. 2017a; Jacobi and Sovinsky 2016; van Ours and Williams 2007; Williams et al. 2011).
Stockpiling may also arise out of fear of future criminalization of the marijuana market. However, the introduction of the tax should likely mitigate this fear, as the tax may be viewed as adding additional legitimacy to the market. Thus, the perceived likelihood that the market will be re-criminalized in the future should fall as a result of the state taxing the marijuana market. In addition, as highlighted in footnote 2, public sentiment is trending in favor of legalization and away from prohibition/criminalization.
In this event study analysis for prices, the week prior to the tax implementation is omitted (i.e., drop the indicator for k = −1). In the event study analysis for quantity, however, the first week of the time window is omitted (i.e., drop the tax indicator for k = −T) to allow for a better visualization of any potential stockpiling effects. Thus the \( \phi_{k} \) coefficients identify treatment effects relative to that omitted week. The figures signify that this zero is imposed, by including a zero point estimate without an associated confidence interval. The baseline results show an eight week window on each side of the tax implementation date, but additional analyses (shown in Appendix Figure 6) are conducted using four and twelve week windows on each side of the tax implementation date.
One could consider estimating a model that includes both the indicators for the weeks before and after the tax and another indicator for the entire post-tax period. Then we could capture the stockpiling effects and the longer run tax incidence all in one regression. Due to the limited post-tax time frame we are able to observe in our data, however, the week indicators and the post-tax variable are highly collinear when implementing this analysis. Thus, while these results are available upon request, we feel that this type of analysis would be better served for an analysis examining a longer post-tax time frame.
An important policy issue is whether this increase pushes consumers above the legal limit for possession. While we are unable to determine the average purchase (in grams) within our data set, some surveys have been conducted to determine average marijuana spending per trip to the dispensary. According to the 2015 Marijuana Business and Investor Survey, the average recreational marijuana consumer in Colorado and Washington spent between $66 and $72 per visit to the dispensary (https://mjbizdaily.com/chart-week-average-marijuana-dispensary-purchase-amounts-range-60-100/). In 2016, a report from Cannabis Market Intelligence Platform Headset (https://www.headset.io/blog/what-does-the-average-cannabis-consumer-look-like) found that most customers spend between $25 and $50 per trip to the dispensary and only 8.2% of customers spent more than $100 per trip. Thus, most consumers are likely purchasing no more than 10 grams in a given trip. As consumers are legally allowed to purchase one ounce (~ 28 g) per trip, even this 15% stockpiling effect is unlikely to push any consumers close to the legal limit for purchasing.
Although not shown in the main text, we also break the Oregon border counties into those that border the Washington recreational market and those that border states without a legal market (i.e., California, Idaho, or Nevada). The main takeaway from this analysis is that the level of stockpiling appears to be much lower in counties that border the states without legal recreational marijuana markets. This lack of a significant stockpiling response may be due to the fact that residents making cross-border purchases in Oregon from these states may not want to stockpile due to legal ramifications of possessing marijuana in states where marijuana possession was still illegal at the time period examined in this study. The full results of this analysis are available upon request.
This finding may also help explain the high R2 we observed in Panel A of Table 3 for non-border counties. The result in Table 3 suggested that there is little time series variation in sales in non-border counties and the results in Fig. 5b suggest that this is the case. Beyond the sharp increase in the week or two before the tax implementation and the immediate drop in sales in the week after the tax implementation, the level of sales in non-border counties shows little variation. Sales are pretty stable around a zero point estimate prior to the stockpiling effect and then are pretty stable around a − 0.1 point estimate in weeks 2–8 after the tax. Given these findings, it may not be all that surprising to see such a large R2 for non-border counties in Table 3 even though there are noticeable changes right around the tax implementation date.
Over the past year, increasing supply in many recreational marijuana markets—and in the Oregon market in particular—has led to substantial decreases in marijuana prices. For more information on these changes in the Oregon market, see https://mjbizdaily.com/decimated-oversupply-oregons-wholesale-marijuana-prices-drop-50-pound/ and https://www.wweek.com/news/2018/04/18/oregon-grew-more-cannabis-than-customers-can-smoke-now-shops-and-farmers-are-left-with-mountains-of-unwanted-bud/.
References
Alm, J., Sennoga, E., & Skidmore, M. (2009). Perfect competition, urbanization, and tax incidence in the retail gasoline market. Economic Inquiry,47(1), 118–134.
Berardi, N., Sevestre, P., Tepaut, M., & Vigneron, A. (2016). The impact of a ‘soda tax’ on prices: Evidence from French micro data. Applied Economics,48(41), 3976–3994.
Bergman, U. M., & Hansen, N. L. (2010). Are excise taxes on beverages fully passed through to prices? The Danish evidence. Working paper.
Besley, T. J., & Rosen, H. S. (1999). Sales taxes and prices: An empirical analysis. National Tax Journal,52(2), 157–178.
Bouga Ela, M. (2018). Legalization of recreational Marijuana in Canada: Implications from provinces’ price difference. Working Paper.
Calonico, S., Cattaneo, M. D., & Titiunik, R. (2014a). Robust nonparametric confidence intervals for regression-discontinuity designs. Econometrica,82(6), 2295–2326.
Calonico, S., Cattaneo, M. D., & Titiunik, R. (2014b). Robust data-driven inference in the regression-discontinuity design. Stata Journal,14(4), 909–946.
Cawley, J., & Frisvold, D. E. (2017). The pass-through of taxes on sugar-sweetened beverages to retail prices: The Case of Berkeley, California. Journal of Policy Analysis and Management,36(2), 303–326.
Chiou, L., & Muehlegger, E. (2008). Crossing the line: direct estimation of cross-border cigarette sales and the effect on tax revenue. The BE Journal of Economic Analysis and Policy,8(1), 2027.
Chiou, L., & Muehlegger, E. (2014). Consumer response to cigarette excise tax changes. National Tax Journal,67(3), 621–651.
Clements, K. W., Lan, Y., & Zhao, X. (2010). The demand for marijuana, tobacco and alcohol: Inter-commodity interactions with uncertainty. Empirical Economics,39(1), 203–239.
Clements, K. W., & Zhao, X. (2009). Economics and marijuana: Consumption. Pricing and Legalisation: Cambridge University Press.
Davis, A. J., Geisler, K. R., & Nichols, M. W. (2016). The price elasticity of marijuana demand: Evidence from crowd-sourced transaction data. Empirical Economics,50(4), 1171–1192.
DeCicca, P., Kenkel, D. S., & Liu, F. (2013). Who pays cigarette taxes? Evidence from consumer search behavior. Review of Economics and Statistics,95, 516–529.
Doye, J., & Samphantharak, M. K. (2008). 2.00 Gas! studying the effects of a gas tax moratorium. Journal of Public Economics,92, 869–884.
Fairbairn, J. W., Liebmann, J. A., & Rowan, M. G. (1976). The stability of cannabis and its preparations on storage. Journal of Pharmacy and Pharmacology,28(1), 1–7.
Fullerton, D., & Metcalf, G. E. (2002). Tax incidence. Handbook of Public Economics,4, 1787–1872.
Grogger, J. (2017). Soda taxes and the prices of sodas and other drinks: Evidence from Mexico. American Journal of Agricultural Economics,99(2), 481–498.
Hansen, B., Miller, K., & Weber, C. (2017a). The taxation of recreational marijuana: Evidence from a reform in Washington State. Working paper.
Hansen, B., Miller, K., & Weber, C. (2017b). Drug trafficking under partial prohibition: Evidence from recreational marijuana (No. w23762). New York: National Bureau of Economic Research.
Hao, Z., & Cowan, B. (2017). The cross-border spillover effects of recreational marijuana legalization (No. w23426). New York: National Bureau of Economic Research.
Harding, M., Leibtag, E., & Lovenheim, M. F. (2012). The heterogeneous geographic and socioeconomic incidence of cigarette taxes: Evidence from Nielsen Homescan Data. American Economic Journal: Economic Policy,4(4), 169–198.
Harris, J. E., Balsa, A. I., & Triunfo, P. (2015). Tobacco control campaign in Uruguay: Impact on smoking cessation during pregnancy and birth weight. Journal of Health Economics,42, 186–196.
Hendel, I., & Nevo, A. (2006a). Measuring the implications of sales and consumer inventory behavior. Econometrica,74(6), 1637–1673.
Hendel, I., & Nevo, A. (2006b). Sales and consumer inventory. The Rand Journal of Economics,37(3), 543–561.
Jacobi, L., & Sovinsky, M. (2016). Marijuana on main street? Estimating demand in markets with limited access. American Economic Review,106(8), 2009–2045.
Kenkel, D. S. (2005). Are alcohol tax hikes fully passed through to prices? Evidence from Alaska. American Economic Review,95(2), 273–277.
Lovenheim, M. F. (2008). How far to the border? The extent and impact of cross-border casual cigarette smuggling. National Tax Journal,61(1), 7–34.
Merriman, D. (2010). The micro-geography of tax avoidance: Evidence from littered cigarette packs in Chicago. American Economic Journal: Economic Policy,2(2), 61–84.
Myles, G. D. (1995). Public economics. Cambridge: Cambridge University Press.
Nicholson, A., Turner, T. M., & Alvarado, E. (2016). Cigarette taxes and cross-border revenue effects: Evidence using retail data. Public Finance Review,44(3), 311–343.
Poterba, J. M. (1996). Retail price reactions to changes in state and local sales taxes. National Tax Journal,49(2), 165–176.
Tremblay, M. J., & Tremblay, V. J. (2017). Tax incidence and demand convexity in Cournot, Bertrand, and Cournot-Bertrand models. Public Finance Review,45(6), 748–770.
van Ours, J. C., & Williams, J. (2007). Cannabis prices and dynamics of cannabis use. Journal of Health Economics,26(3), 578–596.
Wang, E. Y. (2015). The impact of soda taxes on consumer welfare: Implications of storability and taste heterogeneity. The Rand Journal of Economics,46(2), 409–441.
Wang, E., Rojas, C., & Colantuoni, F. (2016). Heterogeneous behavior, obesity, and storability in the demand for soft drinks. American Journal of Agricultural Economics,99(1), 18–33.
Williams, J., van Ours, J. C., & Grossman, M. (2011). Why do some people want to legalize cannabis use? (No. w16795). New York: National Bureau of Economic Research.
Acknowledgements
We would like to thank Scott Akins, Qinglai Meng, and Carol Tremblay for helpful comments and discussion on previous drafts of the paper. We also want to give special thanks to Steve Lurch (State Economist, Washington State), Aaron Hanson (Budget Analyst, Washington Liquor and Cannabis Control Board), Kelly McDermott (Public Records Coordinator, Washington Liquor and Cannabis Control Board), Ramon Cabauatan-Vasquez (Economist, Oregon Health Authority, Public Health Division), and Sam Lambert (Analyst, Benton County) for help in obtaining the data used in this analysis and answering our many questions on the recreational marijuana markets in Oregon and Washington.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Khan, M.S., Thompson, P.N. & Tremblay, V.J. Marijuana tax incidence, stockpiling, and cross-border substitution. Int Tax Public Finance 27, 103–127 (2020). https://doi.org/10.1007/s10797-019-09556-7
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10797-019-09556-7